Urinary creatinine as a biomarker in nutritional and epidemiological studies

Background: Acquiring accurate dietary information is a common problem in nutritional and epidemiological research. To evaluate the validity of dietary assessment methods, biomarkers from 24h urine collections are commonly used. However, obtaining complete collections is a prerequisite for this approach. Aim: To assess the performance of urinary creatinine as a biomarker for identification of incomplete 24h urine collections. Methods: Data was taken from the EFCOVAL project in which 2 non-consecutive 24h urine collections were obtained from 600 subjects (age: 45-65 y) in 5 European countries. Completeness was verified by the ratio of observed to expected creatinine excretion and compared to the recovery of para-aminobenzoic acid (PABA). Cut-offs under which collections were considered incomplete were 0.7 for creatinine and 85% for PABA. Performance was assessed by calculation of sensitivity and specificity. Results: 580 duplicate urine collections were available. For the first and second collection, respectively 7.0% and 11.1% in males (ns) and 7.5% and 14.7% in females (P